Apache Accumulo Random Walk Tests

The randomwalk framework needs to be configured for your Accumulo instance by
doing the following steps:

1.  Make sure you have both ACCUMULO_HOME and HADOOP_HOME set in your 
    $ACCUMULO_HOME/conf/accumulo-env.sh.

2.  Create 'randomwalk.conf' file in the conf directory containing settings
    needed by walkers to connect to Accumulo.

3.  Create a 'walkers' file in the conf directory containing the hostnames of
    the machines where you want random walkers to run.

3.  Create a 'logger.xml' file in the conf directory from logger.xml.example

The command below starts random walkers on all machines listed in 'walkers'.
The argument Image.xml indicates the module to use (which is located at 
conf/modules/Image.xml):

./bin/start-all.sh Image.xml

All modules must be in conf/modules and can be referenced without this prefix.  
For example, a module located at conf/modules/foo/bar.xml is started as
the following:

./bin/start-all.sh foo/bar.xml

This command will load all configuration in the conf directory to HDFS and
start identical random walkers on each node.  These random walkers will
download the current configuration from HDFS and place them in the tmp/
directory.

Random walkers will drop their logs in the logs/ directory.  If you are running
multiple walkers and want error/warns dropped to a NFS hosted logs, please set
NFS_LOGPATH to a NFS-mounted directory and uncomment the NFS appender in logger.xml

You can kill all walkers on the machines listed in the 'walkers' file using
the following command:

./bin/kill-all.sh

-----------------------------------------------------------------------------

Other useful commands:

copy-config.sh		    Copies configuration in conf/ to HDFS

start-local.sh All.xml	    Copies configuration from HDFS into tmp/ and
			    and starts only one local random walker.
 
pkill -f randomwalk.Framework	Stops all local random walkers